Articulatory Text-to-Speech Synthesis Using the Digital Waveguide Mesh Driven by a Deep Neural Network

Amelia Jane Gully; Takenori Yoshimura; Damian Thomas Murphy; Kei Hashimoto; Yoshihiko Nankaku; Keiichi Tokuda

doi:10.21437/Interspeech.2017-900

Articulatory Text-to-Speech Synthesis Using the Digital Waveguide Mesh Driven by a Deep Neural Network

Amelia Jane Gully, Takenori Yoshimura, Damian Thomas Murphy, Kei Hashimoto, Yoshihiko Nankaku, Keiichi Tokuda

Electronic Engineering

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

Abstract

Following recent advances in direct modeling of the speech
waveform using a deep neural network, we propose a novel method that directly estimates a physical model of the vocal tract from the speech waveform, rather than magnetic resonance imaging data. This provides a clear relationship between the model and the size and shape of the vocal tract, offering considerable flexibility in terms of speech characteristics such as age and gender. Initial tests indicate that despite a highly simplified physical model, intelligible synthesized speech is obtained. This illustrates the potential of the combined technique for the control of physical models in general, and hence the generation of more natural-sounding synthetic speech.

Original language	English
Title of host publication	Interspeech 2017
Publisher	ISCA-INST SPEECH COMMUNICATION ASSOC
Pages	234-238
DOIs	https://doi.org/10.21437/Interspeech.2017-900
Publication status	Published - 2017

Publication series

Name	INTERSPEECH
ISSN (Electronic)	1990-9772

Bibliographical note

Keywords

speech synthesis
digital waveguide mesh
deep neural network

Access to Document

10.21437/Interspeech.2017-900Licence: Unspecified

0900
© 2017 ISCA. Uploaded in accordance with the publisher’s self-archiving policy. Further copying may not be permitted; contact the publisher for details
Final published version, 533 KB

http://www.isca-speech.org/archive/Interspeech_2017/pdfs/0900.PDFLicence: Unspecified

JSPS Summer Program Fellowship
Gully, A.
14/06/16 → 24/08/16
Project: Research project (funded) › Research

Cite this

@inproceedings{d5941908105343abbdd75032d2eff15b,

title = "Articulatory Text-to-Speech Synthesis Using the Digital Waveguide Mesh Driven by a Deep Neural Network",

abstract = "Following recent advances in direct modeling of the speechwaveform using a deep neural network, we propose a novel method that directly estimates a physical model of the vocal tract from the speech waveform, rather than magnetic resonance imaging data. This provides a clear relationship between the model and the size and shape of the vocal tract, offering considerable flexibility in terms of speech characteristics such as age and gender. Initial tests indicate that despite a highly simplified physical model, intelligible synthesized speech is obtained. This illustrates the potential of the combined technique for the control of physical models in general, and hence the generation of more natural-sounding synthetic speech.",

keywords = "speech synthesis, digital waveguide mesh , deep neural network",

author = "Gully, {Amelia Jane} and Takenori Yoshimura and Murphy, {Damian Thomas} and Kei Hashimoto and Yoshihiko Nankaku and Keiichi Tokuda",

note = "{\textcopyright} 2017 ISCA. Uploaded in accordance with the publisher{\textquoteright}s self-archiving policy. Further copying may not be permitted; contact the publisher for details",

year = "2017",

doi = "10.21437/Interspeech.2017-900",

language = "English",

series = "INTERSPEECH",

publisher = "ISCA-INST SPEECH COMMUNICATION ASSOC",

pages = "234--238",

booktitle = "Interspeech 2017",

}

TY - GEN

T1 - Articulatory Text-to-Speech Synthesis Using the Digital Waveguide Mesh Driven by a Deep Neural Network

AU - Gully, Amelia Jane

AU - Yoshimura, Takenori

AU - Murphy, Damian Thomas

AU - Hashimoto, Kei

AU - Nankaku, Yoshihiko

AU - Tokuda, Keiichi

PY - 2017

Y1 - 2017

N2 - Following recent advances in direct modeling of the speechwaveform using a deep neural network, we propose a novel method that directly estimates a physical model of the vocal tract from the speech waveform, rather than magnetic resonance imaging data. This provides a clear relationship between the model and the size and shape of the vocal tract, offering considerable flexibility in terms of speech characteristics such as age and gender. Initial tests indicate that despite a highly simplified physical model, intelligible synthesized speech is obtained. This illustrates the potential of the combined technique for the control of physical models in general, and hence the generation of more natural-sounding synthetic speech.

AB - Following recent advances in direct modeling of the speechwaveform using a deep neural network, we propose a novel method that directly estimates a physical model of the vocal tract from the speech waveform, rather than magnetic resonance imaging data. This provides a clear relationship between the model and the size and shape of the vocal tract, offering considerable flexibility in terms of speech characteristics such as age and gender. Initial tests indicate that despite a highly simplified physical model, intelligible synthesized speech is obtained. This illustrates the potential of the combined technique for the control of physical models in general, and hence the generation of more natural-sounding synthetic speech.

KW - speech synthesis

KW - digital waveguide mesh

KW - deep neural network

U2 - 10.21437/Interspeech.2017-900

DO - 10.21437/Interspeech.2017-900

M3 - Conference contribution

T3 - INTERSPEECH

SP - 234

EP - 238

BT - Interspeech 2017

PB - ISCA-INST SPEECH COMMUNICATION ASSOC

ER -

Articulatory Text-to-Speech Synthesis Using the Digital Waveguide Mesh Driven by a Deep Neural Network

Abstract

Publication series

Bibliographical note

Keywords

Access to Document

Projects

JSPS Summer Program Fellowship

Cite this